I need to be able to model a mosfet with values of L, W, Vth, Kprime, lamba and gamma, and I'm trying to figure out how to do this with LTSpice. Looking at ltspice mosfet models, they don't have a separate value for K. They have Cgs, W and L. K is a function of Cgs, W, L, and electron mobility. So I could sort of kluge my own mosfet model by choosing Cgs so that K comes out just right. But to do that I need to know what number ltspice uses for electron mobility. Electron mobility in silicon is somewhere around 1300 cm^2/vs, give or take. What number does ltspice use exactly? Another problem is the body effect parameter, gamma. Of course all the mosfets LTSpice has are three-terminal devices, so they don't have gamma in the list of parameters. You can choose something called nmos4 or pmos4 that has a separate connection for the body, but it doesn't seem to have a model. It has a dialog box where you enter L, W, drain area, source area and some other things that don't get me where I want to go. It doesn't have a model where you can enter parameters like threshhold voltage and K. So that's the biggest roadblock I have. Is it possible to model a mosfet with a separate connection for the body in ltspice. More to the point, does LTSpice even provide a way to take body effect inton account?
ltspice -- body effect and other matters
Started by ●November 5, 2010
Reply by ●November 7, 20102010-11-07
LTSpice support seven different MOSFET models. Your best plan is to read the appropriate section of the manual, then, if necessary, look up the references cited there. IME, mosfet models are a minefield. There was an old thread in here where Jim Thompson and Winfield Hill discussed this very subject in detail, in an attempt to model low level behavior. It got very involved. There may be some material relating to this on Jim's website. Gamma is supprted in the level 1 through level 3 models, BTW. -- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)
Reply by ●November 7, 20102010-11-07
On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse <excretatauris@invalid.invalid> wrote:>LTSpice support seven different MOSFET models. > >Your best plan is to read the appropriate section of the manual, then, if >necessary, look up the references cited there. > >IME, mosfet models are a minefield. There was an old thread in here where >Jim Thompson and Winfield Hill discussed this very subject in detail, in >an attempt to model low level behavior. It got very involved. There may be >some material relating to this on Jim's website. > >Gamma is supprted in the level 1 through level 3 models, BTW.ALL levels of CMOS models attempt (fairly well) to model "body effect" The inaccuracies happen at "sub-threshold". PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure LTspice has an equivalent. The problem is that discrete device manufacturers provide only lower level models, typically Level=3. I do vaguely recall showing that one could fit a Level=7 model to a 2N7002 device. I keep my "Sent" directory trimmed at 2 years, so I can't find the pertinent prior posts. Perhaps someone can find them via GoogleGroups? ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | Ding Dong! The Wicked Old Witch Pelosi is Dead! Hurrah! Hurrah!
Reply by ●November 7, 20102010-11-07
Jim Thompson wrote:>The problem is that discrete device manufacturers >provide only lower level models, typically Level=3. > >I do vaguely recall showing that one could fit a Level=7 model >to a 2N7002 device. > >I keep my "Sent" directory trimmed at 2 years, >so I can't find the pertinent prior posts. >Perhaps someone can find them via GoogleGroups? >http://groups.google.com/group/sci.electronics.design/search?group=sci.electronics.design&q=2N7000+Level-7+Thompson&qt_g=Search+this+group
Reply by ●November 9, 20102010-11-09
On Sun, 07 Nov 2010 09:50:21 -0700, Jim Thompson wrote:> On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse > <excretatauris@invalid.invalid> wrote: > >>LTSpice support seven different MOSFET models. >> >>Your best plan is to read the appropriate section of the manual, then, if >>necessary, look up the references cited there. >> >>IME, mosfet models are a minefield. There was an old thread in here where >>Jim Thompson and Winfield Hill discussed this very subject in detail, in >>an attempt to model low level behavior. It got very involved. There may be >>some material relating to this on Jim's website. >> >>Gamma is supprted in the level 1 through level 3 models, BTW. > > ALL levels of CMOS models attempt (fairly well) to model "body effect" > > The inaccuracies happen at "sub-threshold". > > PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure > LTspice has an equivalent. > > The problem is that discrete device manufacturers provide only lower > level models, typically Level=3. > > I do vaguely recall showing that one could fit a Level=7 model to a > 2N7002 device. > > I keep my "Sent" directory trimmed at 2 years, so I can't find the > pertinent prior posts. > > Perhaps someone can find them via GoogleGroups? >ISTR that some stuff was PDFs posted to a.b.s.e, so they won't have those. I did save copies for a while, but I'm not interested enough in mosfets to have kept them indefinitely. -- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)
Reply by ●November 24, 20102010-11-24
On Sun, 07 Nov 2010 09:50:21 -0700, Jim Thompson wrote:> On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse > <excretatauris@invalid.invalid> wrote: > >>LTSpice support seven different MOSFET models. >> >>Your best plan is to read the appropriate section of the manual, then, if >>necessary, look up the references cited there. >> >>IME, mosfet models are a minefield. There was an old thread in here where >>Jim Thompson and Winfield Hill discussed this very subject in detail, in >>an attempt to model low level behavior. It got very involved. There may be >>some material relating to this on Jim's website. >> >>Gamma is supprted in the level 1 through level 3 models, BTW. > > ALL levels of CMOS models attempt (fairly well) to model "body effect" > > The inaccuracies happen at "sub-threshold". > > PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure > LTspice has an equivalent. > > The problem is that discrete device manufacturers provide only lower > level models, typically Level=3. > > I do vaguely recall showing that one could fit a Level=7 model to a > 2N7002 device. > > I keep my "Sent" directory trimmed at 2 years, so I can't find the > pertinent prior posts. > > Perhaps someone can find them via GoogleGroups? >Trawling around here for something unrelated, I found this, apparently originating from Win Hill, dated march 2005: .SUBCKT 2N7000/PLP 1 2 3 Cgs 2 3 12.3E-12 Cgd1 2 4 27.4E-12 Cgd2 1 4 6E-12 M1 1 2 3 3 MOST1 M2 4 2 1 3 MOST2 D1 3 1 Dbody .MODEL MOST1 NMOS(Level=3 Kp=20.78u W=9.7m L=2u Rs=20m Vto=2 Rd=1.186) .MODEL MOST2 NMOS(VTO=-4.73 Kp=20.78u W=9.7m L=2u Rs=20m) .MODEL Dbody D(Is=125f N=1.023 Rs=1.281 Ikf=18.01 Cjo=46.3p M=.3423 + Vj=.4519 Bv=60 Ibv=10u Tt=161.6n) .ENDS -- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)
Reply by ●November 24, 20102010-11-24
On Wed, 24 Nov 2010 09:22:57 -0800, Fred Abse <excretatauris@invalid.invalid> wrote:>On Sun, 07 Nov 2010 09:50:21 -0700, Jim Thompson wrote: > >> On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse >> <excretatauris@invalid.invalid> wrote: >> >>>LTSpice support seven different MOSFET models. >>> >>>Your best plan is to read the appropriate section of the manual, then, if >>>necessary, look up the references cited there. >>> >>>IME, mosfet models are a minefield. There was an old thread in here where >>>Jim Thompson and Winfield Hill discussed this very subject in detail, in >>>an attempt to model low level behavior. It got very involved. There may be >>>some material relating to this on Jim's website. >>> >>>Gamma is supprted in the level 1 through level 3 models, BTW. >> >> ALL levels of CMOS models attempt (fairly well) to model "body effect" >> >> The inaccuracies happen at "sub-threshold". >> >> PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure >> LTspice has an equivalent. >> >> The problem is that discrete device manufacturers provide only lower >> level models, typically Level=3. >> >> I do vaguely recall showing that one could fit a Level=7 model to a >> 2N7002 device. >> >> I keep my "Sent" directory trimmed at 2 years, so I can't find the >> pertinent prior posts. >> >> Perhaps someone can find them via GoogleGroups? >> > > >Trawling around here for something unrelated, I found this, apparently >originating from Win Hill, dated march 2005: > >.SUBCKT 2N7000/PLP 1 2 3 >Cgs 2 3 12.3E-12 >Cgd1 2 4 27.4E-12 >Cgd2 1 4 6E-12 >M1 1 2 3 3 MOST1 >M2 4 2 1 3 MOST2 >D1 3 1 Dbody >.MODEL MOST1 NMOS(Level=3 Kp=20.78u W=9.7m L=2u Rs=20m Vto=2 Rd=1.186) >.MODEL MOST2 NMOS(VTO=-4.73 Kp=20.78u W=9.7m L=2u Rs=20m) >.MODEL Dbody D(Is=125f N=1.023 Rs=1.281 Ikf=18.01 Cjo=46.3p M=.3423 >+ Vj=.4519 Bv=60 Ibv=10u Tt=161.6n) >.ENDSI did find a simulation comparison on EIGHT different 2N7000 models, only a few came close at sub-threshold. I'll look for the PDF's, and post a link. ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | If Nancy Pelosi gave Obama one of her balls, they'd both have two.
Reply by ●November 24, 20102010-11-24
On Wed, 24 Nov 2010 10:29:25 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:>On Wed, 24 Nov 2010 09:22:57 -0800, Fred Abse ><excretatauris@invalid.invalid> wrote: > >>On Sun, 07 Nov 2010 09:50:21 -0700, Jim Thompson wrote: >> >>> On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse >>> <excretatauris@invalid.invalid> wrote: >>> >>>>LTSpice support seven different MOSFET models. >>>> >>>>Your best plan is to read the appropriate section of the manual, then, if >>>>necessary, look up the references cited there. >>>> >>>>IME, mosfet models are a minefield. There was an old thread in here where >>>>Jim Thompson and Winfield Hill discussed this very subject in detail, in >>>>an attempt to model low level behavior. It got very involved. There may be >>>>some material relating to this on Jim's website. >>>> >>>>Gamma is supprted in the level 1 through level 3 models, BTW. >>> >>> ALL levels of CMOS models attempt (fairly well) to model "body effect" >>> >>> The inaccuracies happen at "sub-threshold". >>> >>> PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure >>> LTspice has an equivalent. >>> >>> The problem is that discrete device manufacturers provide only lower >>> level models, typically Level=3. >>> >>> I do vaguely recall showing that one could fit a Level=7 model to a >>> 2N7002 device. >>> >>> I keep my "Sent" directory trimmed at 2 years, so I can't find the >>> pertinent prior posts. >>> >>> Perhaps someone can find them via GoogleGroups? >>> >> >> >>Trawling around here for something unrelated, I found this, apparently >>originating from Win Hill, dated march 2005: >> >>.SUBCKT 2N7000/PLP 1 2 3 >>Cgs 2 3 12.3E-12 >>Cgd1 2 4 27.4E-12 >>Cgd2 1 4 6E-12 >>M1 1 2 3 3 MOST1 >>M2 4 2 1 3 MOST2 >>D1 3 1 Dbody >>.MODEL MOST1 NMOS(Level=3 Kp=20.78u W=9.7m L=2u Rs=20m Vto=2 Rd=1.186) >>.MODEL MOST2 NMOS(VTO=-4.73 Kp=20.78u W=9.7m L=2u Rs=20m) >>.MODEL Dbody D(Is=125f N=1.023 Rs=1.281 Ikf=18.01 Cjo=46.3p M=.3423 >>+ Vj=.4519 Bv=60 Ibv=10u Tt=161.6n) >>.ENDS > >I did find a simulation comparison on EIGHT different 2N7000 models, >only a few came close at sub-threshold. > >I'll look for the PDF's, and post a link. > > ...Jim ThompsonHere you go... http://analog-innovations.com/SED/Compare2N7000Models.pdf The Supertex model is fairly good, though not perfect, note the "kink". ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | If Nancy Pelosi gave Obama one of her balls, they'd both have two.
Reply by ●November 24, 20102010-11-24
On Wed, 24 Nov 2010 13:09:18 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:>On Wed, 24 Nov 2010 10:29:25 -0700, Jim Thompson ><To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote: > >>On Wed, 24 Nov 2010 09:22:57 -0800, Fred Abse >><excretatauris@invalid.invalid> wrote: >> >>>On Sun, 07 Nov 2010 09:50:21 -0700, Jim Thompson wrote: >>> >>>> On Sun, 07 Nov 2010 08:25:50 -0800, Fred Abse >>>> <excretatauris@invalid.invalid> wrote: >>>> >>>>>LTSpice support seven different MOSFET models. >>>>> >>>>>Your best plan is to read the appropriate section of the manual, then, if >>>>>necessary, look up the references cited there. >>>>> >>>>>IME, mosfet models are a minefield. There was an old thread in here where >>>>>Jim Thompson and Winfield Hill discussed this very subject in detail, in >>>>>an attempt to model low level behavior. It got very involved. There may be >>>>>some material relating to this on Jim's website. >>>>> >>>>>Gamma is supprted in the level 1 through level 3 models, BTW. >>>> >>>> ALL levels of CMOS models attempt (fairly well) to model "body effect" >>>> >>>> The inaccuracies happen at "sub-threshold". >>>> >>>> PSpice has a Level=7 model (BSIM3v3) that works well. And I'm sure >>>> LTspice has an equivalent. >>>> >>>> The problem is that discrete device manufacturers provide only lower >>>> level models, typically Level=3. >>>> >>>> I do vaguely recall showing that one could fit a Level=7 model to a >>>> 2N7002 device. >>>> >>>> I keep my "Sent" directory trimmed at 2 years, so I can't find the >>>> pertinent prior posts. >>>> >>>> Perhaps someone can find them via GoogleGroups? >>>> >>> >>> >>>Trawling around here for something unrelated, I found this, apparently >>>originating from Win Hill, dated march 2005: >>> >>>.SUBCKT 2N7000/PLP 1 2 3 >>>Cgs 2 3 12.3E-12 >>>Cgd1 2 4 27.4E-12 >>>Cgd2 1 4 6E-12 >>>M1 1 2 3 3 MOST1 >>>M2 4 2 1 3 MOST2 >>>D1 3 1 Dbody >>>.MODEL MOST1 NMOS(Level=3 Kp=20.78u W=9.7m L=2u Rs=20m Vto=2 Rd=1.186) >>>.MODEL MOST2 NMOS(VTO=-4.73 Kp=20.78u W=9.7m L=2u Rs=20m) >>>.MODEL Dbody D(Is=125f N=1.023 Rs=1.281 Ikf=18.01 Cjo=46.3p M=.3423 >>>+ Vj=.4519 Bv=60 Ibv=10u Tt=161.6n) >>>.ENDS >> >>I did find a simulation comparison on EIGHT different 2N7000 models, >>only a few came close at sub-threshold. >> >>I'll look for the PDF's, and post a link. >> >> ...Jim Thompson > >Here you go... > >http://analog-innovations.com/SED/Compare2N7000Models.pdf > >The Supertex model is fairly good, though not perfect, note the >"kink". > > ...Jim ThompsonWhen I was searching for this file I ran across an IRF120 test that shows that even the Level=3 model properly show the gate charge plateau. So would could muse if there's a way to take the existing Level=3 stuff and easily convert to Level=7 equivalents, then see what one needs to twiddle to match sub-threshold. ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | If Nancy Pelosi gave Obama one of her balls, they'd both have two.
Reply by ●November 25, 20102010-11-25
On Wed, 24 Nov 2010 13:09:18 -0700, Jim Thompson wrote:> Here you go... > > http://analog-innovations.com/SED/Compare2N7000Models.pdf > > The Supertex model is fairly good, though not perfect, note the "kink".Thanks. (though it was Michael Robinson who originally asked). The Supertex model certainly looks better than the rest. When I get the opportunity, I shall have to curvetrace a real 7000. -- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)